DeepSeek V4 Flash 0731
arcprize.org
DeepSeek V4 Flash 0731
1–10 of 470 posts
Re: DeepSeek V4 Flash 0731
#2results comparable to gpt 5.6 luna but cheaper
promising!
Re: DeepSeek V4 Flash 0731
#3It's always fun when Max reasoning is cheaper than High reasoning.
Re: DeepSeek V4 Flash 0731
#4Kimi K3 was an interesting model only a month ago, and now we're looking at the same performance for 1/20th of the price. Wild how fast this is advancing.
Re: DeepSeek V4 Flash 0731
#5results comparable to gpt 5.6 luna but cheaper promising!
Since the x-axis is log-scaled, DeepSeek is much cheaper than visually implied (mousing over the raw values, it's 1/4th the cost of Luna).
Re: DeepSeek V4 Flash 0731
#6Kimi K3 was an interesting model only a month ago, and now we're looking at the same performance for 1/20th of the price. Wild how fast this is advancing.
Yeah either the benchmark isn't very useful anymore or V4 Flash is a really, really good model.
Re: DeepSeek V4 Flash 0731
#7Price is confounded by VC subsidies, economies of scale, and inference optimizations. I think a more interesting chart would be ARC AGI vs forwards pass flops or ARC AGI vs training tokens. Of course we don't have those numbers for the closed source models or even some of the open weight ones.
Re: DeepSeek V4 Flash 0731
#8It is wild that this a log scale of cost to me!
Re: DeepSeek V4 Flash 0731
#9I've been using it extensively since the release and the best summary I can give is that it's good enough to use it for (almost) everything and cheap enough that the cost are irrelevant. I'm running it in Oh My Pi with a second instance running as "advisor" and even with 5-6 active sessions (effectively 12 streams) I'm struggling to spend more than 5 bucks per day.
OpenCode Go even has double limits temporarily so for 10 USD you effectively get 140 USD of tokens to spend. It would impress me if someone could burn that amount with "normal" usage. Even when running multiple sessions.
I have a Claude Max subscription but I've barely touched it, it just feels like a step back to have to think about limits and usage even though the models are stronger.
The beauty of intelligence at this cost (even if it's not SOTA) is that it opens a whole bunch of new use cases. Test failure in CI? Have the bot automatically propose a fix, its cheap enough that you can discard it w/h issues. Test coverage too low? Auto generate tests on CI for every pull-requests! Monitoring server logs, continuous security audits and investigating every received exception now becomes possible.
I'm thinking about having it automatically filter and re-rank my social media feeds so I can steer the algorithm instead of the other way around.
Perhaps other people (with enormous budgets) were already doing all of the above but for us this is a really exciting release!
Re: DeepSeek V4 Flash 0731
#10Last weeks's discussion (591 points): https://news.ycombinator.com/item?id=49120299